Lessons Learnt from Managing a Case of Dengue Hemorrhagic Fever Complicated with Acute Liver Failure


Dengue is a common arboviral infection with a diverse spectrum of clinical manifestations. Dengue hemorrhagic fever is a more severe form of infection characterized by plasma leak and hemoconcentration. Although hepatic dysfunction is common in dengue illness, massive liver necrosis is rarely reported. Lactic acidosis is a poor prognostic marker in liver failure related to dengue.

Case Report

We report a case of a 43-year-old Sri Lankan Sinhalese woman who presented to our Teaching Hospital, Peradeniya, in the morning with a history of fever, arthralgia, myalgia, and headache of 4 days’ duration. She had been previously diagnosed as having type 2 diabetes mellitus and dyslipidemia for which she received treatment with satisfactory control of the medical conditions and she had normal renal functions.


She had experienced postural dizziness since the afternoon of the previous day and she had had nausea, vomiting, and abdominal pain since the morning of the previous day. Soon after admission she collapsed in our emergency care unit.

She was severely dehydrated with cold clammy peripheries. Her pulse rate was 130 beats per minute with an unrecordable blood pressure. She had reduced breath sounds in the base of her right lung and marked tenderness over the right hypochondrium with flank dullness with shifting. She was drowsy, but arousable.

Her complete blood count showed hemoglobin of 15.3 g/dL, platelet count of 74 × 103/microL, and white cell count of 3.22 × 106/microL. Her serum creatinine was 277 micromole/L with potassium of 5.8 mmol/L. Alanine transaminase (ALT) was 6542 U/L with aspartate transaminase (AST) of 30,617 U/L. Her serum albumin value was 24 g/L.

A bedside ultrasound scan demonstrated bilateral pleural effusions (more in the right) with free fluid in her abdomen. A diagnosis of DHF with decompensated shock complicated with acute liver failure and acute kidney injury was made.

DF was confirmed by positive non-structural protein 1 (NS1) antigen and serotype was identified as dengue virus type 2 (DEN-2). Both dengue immunoglobulin M (IgM) and immunoglobulin G (IgG) were positive suggesting a secondary infection with dengue virus (DENV).

She was screened for alternative causes for liver necrosis including hepatitis A immunoglobulin A, hepatitis B surface antigen, hepatitis C IgM, leptospirosis serology, and rickettsial serology, which were negative. Her baseline transaminase levels and serum creatinine done 1 month earlier at a medical clinic were within the normal range.

She denied taking supra-therapeutic dose of paracetamol or other native medical preparations for her fever. She had been treated with metformin and atorvastatin until the previous night.


She was assumed to be at the peak of the critical phase of DF (that is, 24 hours in the plasma leakage) on admission. On admission her packed cell volume (PCV) was 44%. (Her baseline PCV was 33% for hemoglobin of 10.5 g/dL in her clinic book.) She was given a 10 ml/kg crystalloid bolus (500 ml) over 15 minutes followed by 10 ml/kg bolus over 1 hour.She was given a 10 ml/kg dextran 40% bolus in the next hour. She had heavy per vaginal bleeding and one episode of melena. Her PCV dropped from 44 to 33% without clinical improvement and she had low urine output (< 0.5 ml/kg). She was administered packed cells to maintain the PCV around 40% to a total volume of 1200 ml.During the latter 24 hours of the critical phase of DF, she had severe metabolic acidosis with lactic acidosis: PH of 7.2, bicarbonate 8 mmol/L, and partial pressure of carbon dioxide 16 mmHg with a lactate level of 12 mmol/L, which was corrected medically with 8.4% sodium bicarbonate 200 ml in divided boluses. Her ionized calcium was persistently low and corrected with multiple boluses of intravenously administered calcium gluconate. Her blood sugar was checked every 2 hours and corrected accordingly.At the end of presumed critical phase, we gave her 5500 ml of fluid including normal saline, dextran, and packed cells. She was conscious, rational, but drowsy. She had a spiking high temperature. She was icteric but not pale. She was breathless at rest with oxygen saturation of 85% on room air, which increased to 95% with 60% oxygen via mask.Her pulse rate was 120 beats per minute with blood pressure of 140/100 mmHg. She had bilateral pleural effusions up to mid zone. Her liver was 5 cm below the costal margin with normal upper border and markedly tender. She had gross ascites in a horseshoe-shaped distribution. Her serum creatinine was raised to 345 micromol/L with serum potassium of 5.8 mmol/L and during the last 6 hours of the critical phase she was anuric.Her ALT was 8010 U/L and AST 41546 U/L. Her prothrombin time was 22.1 seconds (control 12 seconds) and activated partial thromboplastin time (APTT) was 42 seconds (control 26 seconds). C-reactive protein (CRP) was 240 U/L. Her blood sugars were elevated toward the end of presumed leaking phase of DHF.At the end of the presumed leaking phase of DHF our patient had: massive liver necrosis; acute kidney injury with acidosis, hyperkalemia, and anuria; deranged clotting with bleeding; and symptomatic volume overload with large plural effusions and gross ascites. Many concerns rose at this point regarding management: The fluid in the third space mainly in the pleural and peritoneal cavities would get reabsorbed and as she probably had an established acute kidney injury with anuria, the reabsorbed fluid would accumulate in her intravascular compartment leading to expansion of intravascular volume and massive volume overload with pulmonary edema and heart failure. Massive liver necrosis with deranged synthetic function would worsen the lactic acidosis which in turn would have a negative effect on the inotropic effect of her heart, clotting derangements might aggravate the bleeding risk, and ongoing hypoxia of the liver might further damage her liver. Sepsis with high fever and elevated inflammatory markers. What is the focus?4.Management of uncontrolled blood sugar. Is it due to her existing type of diabetes or pancreatitis?She was started on continuous renal replacement therapy (CRRT) with CVVHD. We decided to keep CVVHD running and titrate the ultrafiltrate according to the volume state. Observations on central venous pressure (CVP) and blood pressure were made hourly. Intermittent measurements of her inferior vena cava (IVC) diameter and internal jugular vein (IJV) diameter were noted. We assumed that reabsorption of the fluid in the third space would increase the CVP, distend the IJV and IVC, and would increase mainly the diastolic pressure.Depending on the above assumptions, observations were made and ultrafiltrate was gradually increased. Surprisingly, fluid reabsorption occurred in an exponential pattern over a period of 5–6 days and came to a halt abruptly. Maximum ultrafiltrate was 280 ml/hour. Heparin was not used in CVVHD due to high risk of bleeding. She was started on intravenous NAC 100 mg/hour infusion which was continued for 5 days. She was given orally administered metronidazole 400 mg 8 hourly and syrup lactulose to maintain bowel motion 2–3 times per day. She was started on an intravenous infusion of proton pump inhibitors, intravenously administered tranexamic acid, and orally administered norethisterone.She was given intravenously administered vitamin K 10 mg daily for 3 days. She was given 4 units of fresh frozen plasma and 10 units of cryoprecipitate, and 6 units of platelets to correct the coagulopathy. She was transfused with packed cells to maintain PCV around 40% in order to maintain adequate oxygenation of hepatocytes. CVVHD was continued and her lactate level was noted to decline gradually.She had spiking high fevers on day 3 of hospital stay with high CRP. Septic screening was done with blood culture and urine culture and intravenously administered ceftriaxone was changed over to renal-adjusted dose of intravenously administered meropenem and teicoplanin. Later cultures were negative after 72 hours of incubation. However, gradually her fever settled by lysis of fever over the days.Although during the presumed critical phase her blood sugar was rather low, her blood sugar started to rise over the days. She was a type 2 diabetic with good control with metformin. Her blood sugar was checked hourly and insulin infusion was continued and titrated according to her blood sugar. Her amylase was 450 U/L (normal range 1–37 U/L).Over a period of 6 days she was closely monitored. Gradually her transaminases declined, her lactate level normalized, and serum creatinine reduced and normalized. Her urine output gradually increased. CVVHD was terminated after 138 hours of dialysis. She was discharged on ninth day of admission after restoring her full physiology. She was discharged on Mixtard insulin (biphasic isophane insulin injection) for diabetic control. She was reviewed in the ward after 3 days, 7 days, and at 1 month after discharge. Her renal functions and liver functions were within the normal range. She was started again on metformin and atorvastatin 1 week after discharge and insulin was discontinued.


DF is a common mosquito-borne viral disease among humans seen mainly in the Asia-Pacific region. It can present with a diverse clinical spectrum ranging from asymptomatic infection or simple undifferentiated fever to DHF with multiorgan failure.Four distinct dengue viral serotypes (DEN-1 to DEN-4) are known to cause illness. Infection with one serotype confers protection from reinfection with the same serotype, while reinfection with different serotypes confers no long-term protection and may even predispose plasma leak and worse clinical outcome.No specific antiviral therapy is available for DF. Dengue infection can present with various unusual manifestations. Most of these manifestations of DF are under-reported, under-recognized, or not casually linked to DF including hepatitis and liver failure, myositis, and encephalitis and other neurological manifestations.Her DHF was complicated with massive liver necrosis, acute renal failure with anuria and gross volume overload, and secondary sepsis. After an extensive literature search we believe that this case is the first who survived this sort of complicated DF. We would like to discuss the management principles that we employed in this success story.Liver dysfunction is a well-recognized feature in both DF and DHF. Liver involvement in dengue infection could be suspected in patients with DF complaining of abdominal pain, nausea, vomiting, and anorexia.Hepatomegaly is present in both DF and DHF but more common in DF. Clinical jaundice has been detected in 1.7–17% of cases in various series. A mild to moderate increase in the transaminases is common in DF and DHF, and AST was higher than ALT.The AST released from damaged striated muscle, cardiac muscle, and erythrocytes could explain the levels of AST that are higher than those of ALT in patients with DF at an earlier stage. Therefore, a rise in AST might not be a true reflection of hepatic involvement. The pathogenesis of liver injury in dengue infection is yet to be fully elucidated.Possible hypotheses include direct effects of the virus or host immune response on liver cells, circulatory compromise, and metabolic acidosis and/or hypoxia caused by hypotension or localized vascular leakage inside the liver. Studies have shown that DENV readily infects the liver cells in mouse models.High levels of cytokines particularly interleukin-22 (IL-22) and interleukin-17 (IL-17) were found in mouse models which may be responsible for the cytokine-induced liver damage. Sung et al. observed the infiltration of hepatocytes with natural killer cells followed by T cells and this was found to be associated with the apoptosis of hepatocytes.She was started on intravenous NAC infusion at a rate of 100 mg/hour. NAC scavenges free radicals, improves antioxidant defense, and acts as a vasodilator to improve oxygen delivery and consumption.Patients with DHF develop selective plasma leakage manifested as accumulation of fluid in pleural and abdominal cavities and hemoconcentration. The leakage is assumed to last approximately 48 hours and is followed by a spontaneous and rapid resolution but has wide individual variations. Increased vascular permeability is mediated by an interplay between DENV, immune cells and endothelial cells with adhesion molecules, enzymes and cytokines according to the current evidence.Very limited information is available in the literature regarding the rate and duration of plasma leak and about reabsorption of fluid in the pleura and peritoneal cavities during recovery. Understanding the trends in fluid leakage and reabsorption of dengue has been hampered by a lack of animal models.CVP and DBP were considered as hard measurements, whereas IJV and ICV diameters and collapsibility had intra-operator and inter-operator variability. Ultrafiltration of the CVVHD was titrated according to the above measurements. Interesting observations were made during this period.Our patient’s CVP gradually increased reflecting reabsorption of fluid into the vascular compartment and we gradually increased the ultrafiltration starting from a figure of 60 ml/hour. CVP peaked in the 95th hour of CVVHD and ultrafiltration was increased in an exponential manner up to 280 ml/hour. CVP was maintained at the peak for 4 hours and a dramatic decline was noted in the central pressures indicating termination of fluid reabsorption.Fluid dynamics during reabsorption has never been described before in DHF. In the index patient reabsorption took a longer period than is described in a timeframe of 48 hours in the literature and it occurred in an exponential pattern, plateaued, and abruptly ceased.


Dengue hemorrhagic fever Liver failure Acute kidney injury N-acetyl cysteine Packed cell transfusion Continuous veno-venous hemodialysis (CVVHD)

Author : Chamara Dalugama and Indika Bandara Gawarammana